Traffic sign shape classification and localization based on the normalized FFT of the signature of blobs and 2D homographies

摘要: The main goal of a traffic sign recognition system is the detection and every present in scene. Frequently, image processing divided into three parts, namely, segmentation, recognition. In this work, we will focus on block, dividing it two sub-blocks that perform shape classification localization sign, respectively. performed by means signature connected components. Object rotations are tackled with use FFT, normalization object eccentricity improves performance presence projection distortions. effect occlusions lowered removing concave parts shape. Finally, propose novel algorithm, which computes 2D homography, to re-orientate for further steps, like Experimental results, evaluated using huge set randomly generated synthetic images also given, showing great robustness algorithm scaling, rotation, projective deformation, partial noise.